[
https://issues.apache.org/jira/browse/CASSANDRA-20469?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Ariel Weisberg updated CASSANDRA-20469:
---------------------------------------
Bug Category: Parent values: Degradation(12984)
Complexity: Normal
Discovered By: User Report
Fix Version/s: 5.x
Severity: Normal
Status: Open (was: Triage Needed)
> Paxos repair interrupts running transactions
> --------------------------------------------
>
> Key: CASSANDRA-20469
> URL: https://issues.apache.org/jira/browse/CASSANDRA-20469
> Project: Apache Cassandra
> Issue Type: Bug
> Components: Feature/Lightweight Transactions
> Reporter: Ariel Weisberg
> Assignee: Ariel Weisberg
> Priority: Normal
> Fix For: 5.x
>
>
> Noticed that with PaxosV1 there may be transactions being interrupted by auto
> repair on clusters with very low transaction rates and no contention.
> `PaxosCleanupLocalCoordinator` could wait when it encounters a ballot with a
> timestamp that is recent until the transaction timeout has elapsed. This
> would make it impossible to interrupt a transaction that is currently running
> at client unless it is also trying to recover an existing transaction which
> is generally not as common.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
